Research On XML Normalization Based On Semantic Constraints

By virtue of the characteristics of semi-structured, self-description, flexibility and extensibility, XML carries separating structure from content and display into effect, which is available for data exchange between machines, individual display of users and data integration, etc. As a result, XML has been attracting great attention of all circles since it issued and soon becomes the standard of data expression, exchange and integration on Web.With massive XML data emerging on the internet, it presents higher demands for XML. At present, some achievements have appeared in some fields, such as XML data storage and publishing, data query and optimization, and particularly considerable ripe in conversion technology with relational database. Yet such technology just saves structural information rather than evaluates XML database mode from the view of database design, which definitely results in abnormity in XML data processing. In this thesis, we research on XML standardization based on XML semantic restraint, reduce XML data redundancy, keep the data consistence and eliminate XML abnormal operation as well. The main research results are as follows:1. On the basis of definitions of DTD and XML tree, we give the definitions of their paths, nodes and existing corresponding relations among them.2. Bring forward XML functional dependency based on XML tree tuples and, give the conceptions of XML full functional dependency, XML partial functional dependency , XML transitive dependency and XML multivalued dependency as well.3. Give the definitions of logical implication and inference rule for XML functional dependency as well as give proof to the validity and maturity of the inference rule; research on the equivalence of XML functional dependency sets and give the algorithm solving minimum XML functional dependency sets as well.4. Give the definitions of normal form in different levels and, provide corresponding algorithms seeking for XML partial functional dependency pairs and XML transitive dependency groups.5. Give the enhancing rule, rearrangement rule and separating rule which are used to eliminate XML partial functional dependency,XML transitive dependency and XML multivalued dependency respectively; Thereafter, we prove that they all are lossless DTD decomposition preserving XML functional dependencies; After that, we also provide corresponding lossless XML standardization algorithms preserving XML functional dependency.